In conjunction with the recent rapid advances of portable electronic equipment and communications instruments, non-aqueous electrolyte secondary batteries having a high energy density are strongly demanded from the aspects of cost, size and weight reductions. A number of measures are known in the art for increasing the capacity of such non-aqueous electrolyte secondary batteries. For example, JP 3008228 and JP 3242751 disclose negative electrode materials comprising oxides of B, Ti, V, Mn, Co, Fe, Ni, Cr, Nb, and Mo and composite oxides thereof. A negative electrode material comprising M100−xSix wherein x≧50 at % and M=Ni, Fe, Co or Mn is obtained by quenching from the melt (JP 3846661). Other negative electrode materials are known as comprising silicon oxide (JP 2997741), and Si2N2O, Ge2N20 or Sn2N2O (JP 3918311).
Silicon is regarded most promising in attaining the battery's goals of size reduction and capacity enhancement since it exhibits an extraordinarily high theoretical capacity of 4,200 mAh/g as compared with the theoretical capacity 372 mAh/g of carbonaceous materials that are currently used in commercial batteries. Silicon is known to take various forms of different crystalline structure depending on a preparation process. For example, JP 2964732 discloses a lithium ion secondary battery using single crystal silicon as a support for negative electrode active material. JP 3079343 discloses a lithium ion secondary battery using a lithium alloy LixSi (0≦x≦5) with single crystal silicon, polycrystalline silicon or amorphous silicon. Of these, the lithium alloy LixSi with amorphous silicon is preferred, which is prepared by coating crystalline silicon with amorphous silicon resulting from plasma decomposition of monosilane, followed by grinding. However, the negative electrode material therein uses 30 parts of a silicon component and 55 parts of graphite as the conductive agent as described in Example, failing to take full advantage of the potential battery capacity of silicon.
For the purpose of imparting conductivity to negative electrode materials, JP-A 2000-243396 teaches mechanical alloying of a metal oxide such as silicon oxide with graphite and subsequent carbonization; JP-A 2000-215887 mentions coating of Si particles on their surface with a carbon layer by chemical vapor deposition; and JP-A 2002-42806 proposes coating of silicon oxide particles on their surface with a carbon layer by chemical vapor deposition. The provision of particle surfaces with a carbon layer improves conductivity, but is not successful in overcoming the outstanding problems of silicon negative electrodes, i.e., in mitigating substantial volumetric changes associated with charge/discharge cycles or in preventing electricity collection and cycle performance from degrading.
Recently different approaches are thus taken, for example, a method for restraining volume expansion by restricting the percent utilization of silicon battery capacity (JP-A 2000-215887, JP-A 2000-173596, JP 3291260, JP-A 2005-317309), a method of quenching a melt of silicon having alumina added thereto for utilizing grain boundaries in polycrystalline particles as the buffer to volumetric changes (JP-A 2003-109590), polycrystalline particles of mixed phase polycrystals of α- and β-FeSi2 (JP-A 2004-185991), and hot plastic working of a monocrystalline silicon ingot (JP-A 2004-303593).
Means for mitigating volume expansion by tailoring the layer structure of silicon active material are also disclosed, for example, disposition of two layers of silicon negative electrode (JP-A 2005-190902), and coating or encapsulating with carbon or another metal and oxide for restraining particles from spalling off (JP-A 2005-235589, JP-A 2006-216374, JP-A 2006-236684, JP-A 2006-339092, JP 3622629, JP-A 2002-75351, and JP 3622631). In the method of gas phase growing silicon directly on a current collector, degradation of cycle performance due to volume expansion can be restrained by controlling the growth direction (JP-A 2006-338996).
The method of enhancing the cycle performance of negative electrode material by coating silicon surfaces with carbon to be electrically conductive or coating silicon with an amorphous metal layer as mentioned above utilizes only about a half of the silicon's own battery capacity. There is a desire for a higher capacity. As for the polycrystalline silicon having grain boundaries, the disclosed method is difficult to control the cooling rate and hence, to reproduce consistent physical properties.
On the other hand, silicon oxide is represented by SiOx wherein x is slightly greater than the theory of 1 due to oxide coating, and is found on transmission electron microscope analysis to have the structure that crystalline silicon ranging from several to several tens of nanometers is finely dispersed in silicon oxide. The battery capacity of silicon oxide is smaller than that of silicon, but greater than that of carbon by a factor of 5 to 6 on a weight basis. Silicon oxide experiences a relatively less volume expansion. Silicon oxide is thus believed ready for use as the negative electrode active material. Nevertheless, silicon oxide has a substantial irreversible capacity and a very low initial efficiency of about 70%, which requires an extra battery capacity of the positive electrode when a battery is actually fabricated. Then an increase of battery capacity corresponding to the 5 to 6-fold capacity increase per active material weight is not expectable.
The problem of silicon oxide to be overcome prior to practical use is a substantially low initial efficiency. This may be overcome by making up the irreversible fraction of capacity or by restraining the irreversible capacity. The method of making up the irreversible fraction of capacity by previously doping silicon oxide with Li metal is reported effective. Doping of lithium metal may be carried out by attaching a lithium foil to a surface of negative electrode active material (JP-A 11-086847) or by vapor depositing lithium on a surface of negative electrode active material (JP-A 2007-122992). As for the attachment of a lithium foil, a thin lithium foil that matches with the initial efficiency of silicon oxide negative electrode is hardly available or prohibitively expensive if available. The deposition of lithium vapor makes the fabrication process complex and is impractical.
Aside from lithium doping, it is also disclosed to enhance the initial efficiency of negative electrode by increasing a weight proportion of silicon. One method is by adding silicon particles to silicon oxide particles to reduce the weight proportion of silicon oxide (JP 3982230). In another method, silicon vapor is generated and precipitated in the same stage as is produced silicon oxide, obtaining mixed solids of silicon and silicon oxide (JP-A 2007-290919). Silicon as an active material has both a high initial efficiency and a high battery capacity as compared with silicon oxide, but displays a percent volume expansion as high as 400% upon charging. Even when silicon is added to a mixture of silicon oxide and carbonaceous material, the percent volume expansion of silicon oxide is not maintained, and eventually at least 20 wt % of carbonaceous material must be added in order to suppress the battery capacity at 1,000 mAh/g. The method of obtaining the mixed solids by simultaneously generating silicon and silicon oxide vapors suffers from the working problem that the low vapor pressure of silicon necessitates the process at a high temperature in excess of 2,000° C.
As discussed above, the silicon-based active material still has a problem to be solved prior to practical use, independent of whether it is based on a metal element or an oxide thereof. There is a desire to have a negative electrode active material which can restrain the volumetric change associated with occlusion and release of lithium, mitigate a lowering of conductivity due to atomization by fissure of particles and separation of particles from the current collector, be manufactured on a mass scale at a low cost, and comply with the application as in mobile phones where repetitive cycle performance is of high priority.